[U-Boot] [U-Boot,v2,4/7] mach-snapdragon: Fix UART clock flow

[Jorge Ramirez-Ortiz <jramirez@baylibre.com>]

On 05/28/2018 10:25 PM, Ramon Fried wrote:
> On Mon, May 28, 2018 at 11:14 PM, Ramon Fried <ramon.fried@gmail.com> wrote:
>> On Mon, May 28, 2018 at 11:07 PM, Jorge Ramirez-Ortiz
>> <jramirez@baylibre.com> wrote:
>>> On 05/28/2018 10:01 PM, Ramon Fried wrote:
>>>> On Mon, May 28, 2018 at 10:59 PM, Jorge Ramirez-Ortiz
>>>> <jramirez@baylibre.com> wrote:
>>>>> On 05/28/2018 09:48 PM, Ramon Fried wrote:
>>>>>> On Mon, May 28, 2018 at 10:24 PM, Ramon Fried <ramon.fried@gmail.com>
>>>>>> wrote:
>>>>>>> On Mon, May 28, 2018 at 10:19 PM, Jorge Ramirez-Ortiz
>>>>>>> <jramirez@baylibre.com> wrote:
>>>>>>>> On 05/28/2018 09:12 PM, Tom Rini wrote:
>>>>>>>>> On Wed, May 16, 2018 at 12:13:39PM +0300, Ramon Fried wrote:
>>>>>>>>>
>>>>>>>>>> UART clock enabling flow was wrong.
>>>>>>>>>> Changed the flow according to downstream implementation in LK.
>>>>>>>>>>
>>>>>>>>>> Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
>>>>>>>>> Applied to u-boot/master, thanks!
>>>>>>>>>
>>>>>>>> Ramon, did you re-test this one on the 820 as we discussed?
>>>>>>>> Sorry Tom, when I tested this on Friday it broke my 820 (I should have
>>>>>>>> reported it to the ML).
>>>>>>>>
>>>>>>>> I think it introduces a regression but I'll let Ramon to confirm.
>>>>>>> Hi.
>>>>>>> It's funny, I'm debugging it now. don't have any conclusions yet but I
>>>>>>> was under the assumption that it won't get merged as it was
>>>>>>> missing Reviewed-by.
>>>>>>> Let me get back to you on these one in couple of hours.
>>>>>>> Thanks,
>>>>>>> Ramon.
>>>>>> I just toasted my 820 board...:(
>>>>>> I can only get a replacement by Thursday.
>>>>>> I can't find any explanation why the 820 should be affected, as the
>>>>>> clock configuration for it is empty.
>>>>>> and it's pre-initalized by the uart.
>>>>>> Jorge, you previously tested from my branch, care to test from master
>>>>>> to see if it's working ?
>>>>>
>>>>> so just pull your master branch again and retest?
>>>> not mine. upstream master.
>>>
>>> 2 days ago ffada23 db820c: set clk node to be probed before relocation
>>> Ramon Fr..[Tom Rini]
>>> 2 days ago 0ed34aa Remove CONFIG_MVGBE from config_whitelist.txt
>>> Chris Pa..[Tom Rini]
>>> 2 days ago 5ce9aca PCI: Document pciauto_region_allocate()
>>> Tuomas T..[Tom Rini]
>>> 2 days ago d71975a PCI: autoconfig: Don't allocate 64-bit addresses to
>>> 32-bit.. Tuomas T..[Tom Rini]
>>> 2 days ago ed12a89 PCI: Add newlines to debug prints in pci_auto_common.c
>>> Tuomas T..[Tom Rini]
>>>
>>> the one on top is the last commit that works on the 820 from the current
>>> uboot master branch.
>> Thanks Jorge.
>> I see where the problem is. there's no clock defintions in 820 device tree,
>> so msm_uart_clk_init() fails. I'll add dummy definitions meanwhile and
>> after I'll finish
>> with 410, I'll work on the correct clock values for 820. nevertheless
>> the UART will work as
>> LK already initalizes it with the right values.
>> expect a fix patch shortly.
>>
>> Thanks,
>> Ramon.
> Jorge, I just sent you a fix. can you test it and if it works I'll
> push it upstream.

yes I can see the console now so that fix is good.

however there must be some other regression lurking because the system 
wont boot a kernel from the SD card